feat(docs): add more configuration details to the HPA docs

[lc525]

• improve overall docs clarity
• explain prometheus-adapter ConfigMap customisation options
• mention that spec.replicas needs to be present in the Model CR for HPA to work
• change apply to replace in kubectl command updating the prometheus-adapter ConfigMap
• clarify in what namespaces we should load the various manifests in the example
• add cluster operation guidelines for HPA-based autoscaling

[paulb-seldon]

Should we link to this page so users understand the metrics Core 2 makes available to them (perhaps in the Installing and configuring the Prometheus Adapter) section?

[paulb-seldon]

Consider renaming the section titled Seldon Core 2 scheduler interactions with HPA-based scaling? I think this is an important section, but I wouldn't know from the title that there is is relevant operational information there related to resource states, handling spikes, tips on configuration... perhaps something like Operational Guidelines for coordinating HPA Scaling and Core 2 Scheduling?

[paulb-seldon]

I agree that this is long, however I do think we should keep it as is for now. Thankfully, there is the right side nav that outlines the high level sections.

I would recommend perhaps simplifying the the titles of those sections if possible, and then I think that's enough for a user to have bearings when going through these docs. As we refactor IA of the docs more holistically, I think we will have to re-review content and how we break it up.

[paulb-seldon]

Overall LGTM. Have made some minor edits and possible suggestions in comments